ORIGINAL RESEARCH
published: 20 April 2015
Edited by:
Senjie Lin,
University of Connecticut, USA
Reviewed by:
Tatiana Rynearson,
University of Rhode Island, USA
Tom O. Delmont,
Marine Biological Lab, USA
*Correspondence:
Zhenfeng Liu,
Department of Biological Sciences,
University of Southern California, 3616
Trousdale Parkway, AHF 231,
Los Angeles, CA 90089-0371, USA
zhenfeng.liu1@gmail.com
Specialty section:
This article was submitted to
Aquatic Microbiology,
a section of the journal
Frontiers in Microbiology
Received: 05 December 2014
Accepted: 30 March 2015
Published: 20 April 2015
Citation:
Liu Z, Jones AC, Campbell V,
Hambright KD, Heidelberg KB and
Caron DA (2015) Gene expression in
the mixotrophic prymnesiophyte,
Prymnesium parvum, responds to
prey availability.
Front. Microbiol. 6:319.
doi: 10.3389/fmicb.2015.00319
Frontiers in Microbiology | www.frontiersin.org
doi: 10.3389/fmicb.2015.00319
Gene expression in the mixotrophic
prymnesiophyte, Prymnesium
parvum, responds to prey availability
Zhenfeng Liu 1*, Adriane C. Jones 1 , Victoria Campbell 1 , K. David Hambright 2 ,
Karla B. Heidelberg 1 and David A. Caron 1
1
Department of Biological Sciences, University of Southern California, Los Angeles, CA, USA, 2 Program in Ecology and
Evolutionary Biology, Department of Biology, University of Oklahoma, Norman, OK, USA
The mixotrophic prymnesiophyte, Prymnesium parvum, is a widely distributed alga with
significant ecological importance. It produces toxins and can form ecosystem disruptive
blooms that result in fish kills and changes in planktonic food web structure. However, the
relationship between P. parvum and its prey on the molecular level is poorly understood.
In this study, we used RNA-Seq technology to study changes in gene transcription of
P. parvum in three treatments with different microbial populations available as potential
prey: axenic P. parvum (no prey), bacterized P. paruvm, and axenic P. parvum with
ciliates added as prey. Thousands of genes were differentially expressed among the
three treatments. Most notably, transcriptome data indicated that P. parvum obtained
organic carbon, including fatty acids, from both bacteria and ciliate prey for energy and
cellular building blocks. The data also suggested that different prey provided P. parvum
with macro- and micro-nutrients, namely organic nitrogen in the form of amino acids
from ciliates, and iron from bacteria. However, both transcriptomic data and growth
experiments indicated that P. parvum did not grow faster in the presence of prey despite
the gains in nutrients, although algal abundances attained in culture were slightly greater
in the presence of prey. The relationship between phototrophy, heterotrophy and growth
of P. parvum is discussed.
Keywords: Prymnesium parvum, transcriptomics, gene expression, mixotrophic protist, golden algae, harmful
algae
Introduction
Prymnesium parvum is a globally distributed prymnesiophyte alga of considerable ecological
importance. P. parvum occurs in ecosystems across a wide range of salinities and environmental
conditions (Guo et al., 1996; Barkoh and Fries, 2010; Johnsen et al., 2010), possesses a complex life
history including haploid and diploid stages (Larsen, 1999), and exhibits mixtrophic abilities (Till-
mann, 2003). The combined photosynthetic and heterotrophic capabilities of P. parvum enable
the alga to form Ecosystem Disruptive Algal Blooms (EDABS) (Sunda et al., 2006). The disruptive
nature of the alga to the ecosystems in which it blooms has implicated P. parvum as a cause of
significant economic loss to the aquaculture industry worldwide (Barkoh and Fries, 2010).
Commonly referred to as the “golden alga,” P. parvum has appeared and attained bloom abun-
dances in several waterways of the southwestern United States during the last two decades (Roelke
et al., 2011; Hambright et al., 2015). Genetic analyses of P. parvum populations have indicated
1 April 2015 | Volume 6 | Article 319
Liu et al.
a relatively recent appearance of this species in the region,
perhaps a consequence of a recent introduction of a strain
from Europe, and subsequent range expansion (Lutz-Carrillo
et al., 2010; Hambright et al., 2015). The alga has resulted
in ecosystem disturbances ranging from changes in plank-
ton food web structure to massive fish kills (Southard
et al., 2010; Jones et al., 2013). Various reports have noted
that substances produced by P. parvum can immobilize
a range of co-occurring species and thereby aid in the
capture of prey that are often considerably larger than
the alga (Henrikson et al., 2010; Remmel and Hambright, 2012).
These substances may also act as predator deterrents for potential
consumers or allelopathic agents retarding the growth of com-
peting phytoplankton (Schmidt and Hansen, 2001; Fistarol et al.,
2003; Granéli and Salomon, 2010; Remmel et al., 2011).
P. parvum exhibits robust phototrophic growth in culture,
a characteristic that is not always common for mixotrophic
microalgae (Jones, 1997). Aside from a requirement for the vita-
mins thiamine (B1 ) and cobalamin (B12 ) (McLaughlin, 1958;
Rahat and Reich, 1963), the alga is capable of growth on inor-
ganic medium. Maximal growth rates of 0.6–0.8 d−1 have been
reported for Prymnesium species in axenic culture (Granéli et al.,
2012; Hambright et al., 2014). Conversely, growth of P. parvum
in the dark is not supported by the presence of prey, and has only
been achieved when the growth medium has been supplemented
with high concentrations of specific organic compounds (most
notably glycerol) (Rahat and Jahn, 1965). Moreover, growth rates
in the dark in the presence of high concentrations of glycerol that
have been reported for this species are considerably slower than
growth in the light (Rahat and Jahn, 1965).
Despite a very limited ability to grow exclusively heterotroph-
ically, P. parvum grown in the light displays clear predatory
behaviors, and has been shown to be predatory on a number
of types of microbes including bacteria, protists and a variety
of metazoan zooplankton (Tillmann, 2003; Sopanen et al., 2006;
Brooks et al., 2010). It has been suggested that up to 78% of
the cellular nitrogen and 45% of the cellular phosphorus may be
acquired via predation (Carvalho and Granéli, 2010). The het-
erotrophic capabilities of the alga include the production of a
suite of toxins that exhibit hemolytic, ichthyotoxic, cytolytic, and
potentially neurotoxic activities (Edvardsen and Paasche, 1998;
Barkoh and Fries, 2010; Henrikson et al., 2010; Bertin et al., 2012;
Granéli et al., 2012), and toxin characterization is an area of
active research (Schug et al., 2010; Weissbach and Legrand, 2012;
Manning and La Claire, 2013).
An alternative approach to understanding the physiological
response of microbial eukaryotes to environmental factors has
recently been enabled through genomic and transcriptomic anal-
ysis (Bik and Thomas, 2012; Murray et al., 2012). The latter
approach has proven particularly fruitful for microbial eukary-
otes in recent years (Keeling et al., 2014) because the larger
genomes of many of these species relative to bacteria and archaea
incur much greater sequencing costs and difficulties with genome
assembly. The genome of P. parvum has not yet been sequenced,
but a few studies examining gene expression in the species using
a variety of methods have been conducted. La Claire (2006)
established the first EST database for P. parvum and examined
the most frequently detected genes. Beszteri et al. (2012) used
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
the same approach to study P. parvum gene expression under
N or P limitations. Freitag et al. (2011) correlated the expres-
sion of certain polyketide synthase genes (genes assumed to play
a role in toxin production) by qPCR with toxicity of the alga
when cultures were exposed to physiological shock. Koid et al.
(2014) compared the gene contents of P. parvum and other three
prymnesiophytes on a whole transcriptome level. These studies
have provided baseline information on gene expression by the
prymnesiophyte.
We conducted a transcriptomic analysis of P. parvum using
RNA-Seq designed to specifically examine the effect of the pres-
ence or absence of prey on gene expression. Gene expression of
the alga was analyzed when grown in laboratory culture for 2–3
weeks in three situations; (1) axenically, (2) in the presence of
a mixed bacterial assemblage (potential prey and/or symbiotic
interactions), and (3) grown in the presence of a ciliated proto-
zoan that was readily consumed by the prymnesiophyte. Gene
expression of P. parvum was significantly affected by the pres-
ence of other microbial taxa. Transcription data suggested that
P. parvum obtained organic carbon from both the bacteria and
ciliate populations. The data also suggested that the two differ-
ent prey types provided different nutrients to P. parvum, namely
organic nitrogen from the ciliate and iron from the bacteria.
However, despite providing nutrients, the presence of prey had
little positive impact on the growth rate of P. parvum, although
nutrients released from prey maintained algal population growth
longer than in axenic cultures.
Materials and Methods
Organisms and Cultures
Prymnesium parvum strain UOBS-LP0109 (Texoma1) was iso-
lated from Lake Texoma, Oklahoma, USA. It was subsequently
made axenic by micropipetting single cells through rinses of
sterile medium, and then cultured in sterile L1 medium with-
out silica (https://ncma.bigelow.org). The bacterivorous ciliate,
Uronema marina, was isolated from Buzzards Bay, Massachusetts
and maintained on its attendant bacterial flora by periodic sub-
culturing into filtered sterile seawater containing 0.01–0.001%
yeast extract.
Transcriptomics Experiment
Three treatments were performed for transcriptomic analysis:
axenic P. parvum, non-axenic P. parvum with its attendant bac-
terial flora (hereafter called bacterized treatment), and an axenic
culture of P. parvum to which an aliquot of the bacterized ciliate
culture was added (hereafter called ciliate treatment). The latter
treatment was prepared by inoculating an axenic P. parvum cul-
ture in mid-late exponential phase with an aliquot of the ciliate
culture in early stationary phase (∼50,000 ciliates· ml−1 ) to yield
an alga:ciliate ratio of 100:1. Ciliate biomass strongly dominated
the ciliate culture at the time of innoculation, although attendant
bacteria were also present at ∼105 · ml−1 in the combined cul-
ture. Cultures were examined using a dissecting microscope every
30–40 min following addition of the ciliates to determine when
ciliates were eliminated by the alga. All ciliates were rapidly killed
(<6 h) after addition of ciliates to the P. parvum culture, and the
cultures were harvested as soon as no live ciliates remained.
2 April 2015 | Volume 6 | Article 319
Liu et al.
All three treatments were grown in 1–2 L volumes in 2800 ml
Pyrex glass Fernbach flasks in L1 medium minus silica at 18
ppt salinity and 18◦ C. Cultures were incubated at an irradiance
of ∼300 µE·m−2 s−1 , measured using a QSL-100 sensor with
QSP-170 deckbox (Biospherical Instruments Inc.). Cultures were
incubated on a 12:12 h light:dark cycle.
Growth Experiment
The effect of the presence or absence of prey on the growth of
P. parvum was established in batch cultures of the alga. All cul-
tures used in growth experiments were grown under the same
conditions as the transcriptome experiment except that they were
in ∼100 mL volumes. The same three treatments of P. parvum
that were used for transcriptomic analysis were examined. All
treatments were conducted in duplicate. On day 16, cells from
a ciliate culture were added to an axenic P. parvum culture at a
ratio of 100:1 (alga:ciliate). Growth of P. parvum was monitored
by performing direct microscopic cell counts on samples peri-
odically removed from each flask over a 28-day period. Growth
rates in different treatments were compared using One-Way
ANOVA test.
RNA Extraction and cDNA Production
All samples were harvested during mid-exponential growth
phase in the middle of the 12 h light cycle. Cultures were spun
down in an Eppendorf 5810R centrifuge using the A-4-62 rotor
at 3200 rcf for 15 min at 15◦ C. The supernatant was carefully
decanted, and 1–2 ml of TRI reagent (Ambion) was added to the
pellet and vortexed until the pellet fully dissolved. Homogenates
were then either processed immediately using Ribopure kit
(Ambion), or stored at −80◦ C for later processing. Eluted RNA
was treated with DNase (Sigma) to remove DNA contamination.
Samples were cleaned and concentrated using RNA Clean and
Concentrator-25 (Zymo Research). RNA was quantified using a
Qubit 2.0 Fluorometer (Invitrogen) and run on an E-gel iBase
with E-gel Gel EX 1% (Invitrogen) to check for nucleic acid
quality.
RNA samples were sent to the National Center for Genome
Resources for processing and sequencing. RNA was quantified
again at the sequencing center using Invitrogen Qubit Q32855,
and RNA quality was assessed using the Agilent 2100 Bioan-
alyzer. cDNA libraries were made from 2 µg RNA using Illu-
mina’s TruSeq RNA Sample Preparation Kit. The average insert
size of each library ranged from 250 to 350 bp. Libraries were
sequenced on an Illumina HiSeq 2000 which generated 2 ×
50 bp (paired-end) reads. The original sequence data are pub-
licly available from NCBI Sequence Read Archive under acces-
sion number SRA166613 and sample IDs MMETSP0006_2,
MMETSP0815, and MMETSP0008_2, for axenic, bacterized and
ciliate treatment, respectively.
Sequencing and Bioinformatic Analyses
Bioinformatic analysis procedures were adapted and devel-
oped from guidelines established by the Marine Microbial
Eukaryote Transcriptome Sequencing Project (https://www.
marinemicroeukaryotes.org) (Keeling et al., 2014). Sequences of
all treatments were first checked for quality using the FASTX
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
toolkit (http://hannonlab.cshl.edu/fastx_toolkit/index.html)
with options “-p 80 -q 20.”
An assembled transcriptome of P. parvum was then gener-
ated as a reference for downstream alignment processes. All
sequences of the ciliate treatment were excluded from the assem-
bly to ensure that the P. parvum assembled transcriptome was not
contaminated with ciliate sequences that might have remained
after ciliates were killed by the alga. Although the bacterized treat-
ment and ciliate treatment also contained bacteria, the sequence
data from those treatments should have been reasonably free of
bacterial sequences because RNA was polyA-selected in our pro-
tocol for all treatments. Sequences from the axenic and bacterized
treatments were combined and assembled de novo using a hybrid
approach containing both a de Bruijn graph and an overlap-based
algorithm method. Sequences were first assembled using ABySS
v. 1.3.2 (Simpson et al., 2009) at four different k-mer settings of
19, 25, 31, and 37. The resulting four assemblies were merged
using Trans-ABySS v. 1.4.4 (Robertson et al., 2010). Redundant
contigs were removed from the assembly using CD-Hit-EST v.
4.5.7 (Li and Godzik, 2006) with option “-c 0.99.” The remain-
ing contigs were then further assembled using the overlap-based
program CAP3 (Huang and Madan, 1999) with options “-p 99 -o
50 −k 0.” Scaffolding of the resulting contigs was inferred using
ABySS with default parameters (Simpson et al., 2009). The Gap-
Closer v. 1.12 application of the SOAPdenovo (Luo et al., 2012)
was employed to try to fill in gaps in the assembly. Scaffolds
were broken into contigs where gaps remained unfilled. Contigs
shorter than 150 bp were discarded. CD-Hit-EST was used again
to remove redundant contigs.
All contigs were searched against SILVA database (Quast et al.,
2013) using BLAST to identify and remove rRNA sequences. Pro-
tein coding genes with at least 150 bp were predicted from the
remaining contigs using ESTscan v. 3.0.3 (Iseli et al., 1999). Genes
were annotated using an evalue cutoff of 1e-5 based on a vari-
ety of database searches including HMMER3 v. 3.1b1 (http://
hmmer.janelia.org) searches against Pfam (http://pfam.sanger.
ac.uk) and Tigrfam (http://www.jcvi.org/cgi-bin/tigrfams/index.
cgi) databases. A BLAST search against NCBI nr database was
also carried out using the same e-value cutoff. Genes without
hits to Pfam or Tigrfam were annotated based on their best hits
in the NCBI nr database. KEGG and Gene Ontology annota-
tions were also obtained using KEGG Automatic Annotation
Server (http://www.genome.jp/tools/kaas/) and Blast2GO v. 2.5
(Conesa et al., 2005), respectively, to provide additional func-
tional understanding of the genes. Annotations of some genes of
interest were edited after manual inspection of all the information
available.
Sequences from each of the three datasets were aligned back
to the assembled transcriptome using BWA v. 0.6.1-r104 (Li and
Durbin, 2009). Read pairs that were aligned to genes correctly
(i.e., paired reads aligned to the same gene, on the opposite
strand, with fragment size of ∼200 bp) were counted using a cus-
tom PERL script. Statistical analyses of the read counts of each
gene were carried out using the “exact test” function of edgeR
v. 3.6.2 (Robinson et al., 2010) with common dispersion set at
0.1. Pairwise comparisons between all three treatments were car-
ried out. P-values were adjusted to false discovery rate using
3 April 2015 | Volume 6 | Article 319
Liu et al.
p.adjust in R software v. 3.1.0 (Benjamini and Hochberg, 1995).
Only genes with adjusted p-values smaller than 0.05 were deemed
as having significantly different expression levels between treat-
ments. Normalized expression levels in the form of FPKM (Frag-
ments per kilobase exon per million fragments aligned) values
were calculated. Expression levels in bacterized and ciliate treat-
ments relative to those in axenic treatment were calculated and
plotted in figures.
Results
Growth Experiment
A growth experiment was conducted to determine whether the
presence of prey (bacteria or ciliates) had a significant effect on
the growth rate or overall yield of P. parvum (Figure 1). Growth
rates of the alga were not significantly different (p = 0.20,
ANOVA) among the three treatments in early exponential phase
(0.36 ± 0.02, 0.42 ± 0.01, 0.38 ± 0.04 d−1 during the first 9 days
of the experiment for axenic, bacterized, and ciliate treatment,
respectively. Axenic and ciliate treatments did not differ during
that period). The addition of ciliates at late exponential phase
of the alga on day 16 allowed P. parvum to continue growing
(0.12 ± 0.05 d−1 ) while cultures without ciliates exhibited neg-
ligible growth (0.03 ± 0.01 d−1 ). P. parvum in the bacterized
treatment reached the maximum abundance of 4.28 ± 0.03 × 106
cells mL−1 , which is slightly higher than that of the axenic treat-
ment (3.80 ± 0.17 × 106 cells mL−1 ). The addition of ciliates
also increased the maximum abundance of P. parvum to (4.95
± 0.25 × 106 cells mL−1 ).
FIGURE 1 | Growth of the prymnesiophyte, Prymnesium parvum, in
axenic culture (circles, solid line), in non-axenic culture with attendant
bacterial flora (squares, dashed line), and in axenic culture with the
addition of ciliates at late exponential phase (triangles, dotted line).
Arrow indicates Day 16 when ciliates were added to the axenic P. parvum
culture. In the transcriptome experiment, cells were harvested from axenic and
bacterized treatment just prior to the addition of ciliates, while cells were
harvested from ciliate treatment ∼6 h after the addition of ciliates. Symbols and
error bars represent mean ± standard deviation of two replicates.
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
Overview of the Transcriptome
The assembled transcriptome of P. parvum contained 55,984
contigs and 43,058 genes or partial genes, totaling 47.42 mil-
lion base pairs (Table 1). Most of the physiologically necessary
genes were found in the transcriptome (Table S1). For example,
the transcriptome contained all genes necessary for glycolysis,
the TCA cycle, nitrogen metabolism, nucleotide biosynthesis and
almost all genes required for the biosynthesis of all 20 amino
acids. Based on this information, it is reasonable to assume that
the transcriptome had enough sequencing depth to cover most
of the transcribed genes. Consistent with previous experiments
(McLaughlin, 1958; Rahat and Reich, 1963), the P. parvum tran-
scriptome contained no genes for the synthesis of thiamine and
cobalamin but did contain most of the genes required for the
synthesis of other vitamins.
Among the 43,058 genes identified in this study, only a
small percentage (9.2% or 3972 genes) had significantly different
expression levels between at least a pair of treatments (Figure 2).
Similar numbers of genes (1600∼1900) had higher expression
levels in each of the three treatments (1612, 1861, and 1645
genes in the axenic, bacteria and ciliate treatments, respectively;
Figure 2). The overlap among those sets of genes, however, was
quite interesting. A total of 618 genes, or 33 and 38% of all
genes with elevated expression levels in the bacterized and cili-
ate treatments, respectively, had higher expression levels in both
treatments compared to the axenic treatment. In comparison,
426 genes were shared between the axenic and bacterized treat-
ments, and only 102 genes were shared between the axenic and
ciliate treatments, in terms of genes with higher expression levels
compared to the third treatment (Figure 2).
A more detailed evaluation of differential gene expression and
function with respect to roles in P. parvum showed that most
(>50%) of its genes had no matches or functional annotations
to existing databases, owing to the relative scarcity of protistan
sequence data (Caron et al., 2009). Nevertheless, we were able
TABLE 1 | Summary of three transcriptome datasets of Prymnesium
parvum.
Axenic Bacterized Ciliate
No. of readsa 19,277,859 pairs 33,615,796 pairs 17,560,521 pairs
Size of assembled 55,984 contigs; 47.42 Mbp; N50 = N/A
transcriptomeb 1281 bp; 43,058 predicted genes;
average 800 bp
Reads mapped 77.6% 75.8% 69.3%
back to
transcriptiome
Reads mapped 61.5% 62.0% 61.0%
back to predicted
genes
a Afterquality filtering (see Materials and Methods for quality filtering standards).
b The transcriptome of axenic and bacterized treatments were combined to generate
an assembly. The ciliate treatment was not included in the assembly to exclude cili-
ate sequences from the assembled transcriptome. Only contigs and genes with at least
150 bp were retained for further analysis.
4 April 2015 | Volume 6 | Article 319
Liu et al.
FIGURE 2 | Distribution of genes according to changes in their
expression levels among three treatments. Numbers in each area indicate
numbers of genes whose expression levels were higher in that treatment
compared to the other two or in two treatments compared to the third.
to infer interesting patterns from changes in the transcriptional
patterns of those genes with putative functions (Figure 3). Genes
belonging to the same pathway or metabolic function had very
similar transcriptional patterns among the treatments. For exam-
ple, all differential expressed genes associated with fatty acid
metabolism had higher expression levels in the presence of either
prey type and formed a cluster in the upper right quadrant
of Figure 3, while genes associated with iron uptake clustered
together in the lower left quadrant (i.e., the latter genes had lower
expression levels in the presence of both prey). Figure 3 also illus-
trates that genes associated with different metabolic functions
often demonstrated different transcriptional patterns (as noted
above). While changes in transcription do not necessarily equate
with protein concentrations or enzymatic activities, the tran-
scriptional patterns observed in this study provided insights into
the relationship between P. parvum and the presence of potential
prey, and different prey types.
Carbon Metabolism
Genes involved in several carbon metabolism pathways had
higher expression levels in both treatments with prey, compared
to the axenic treatment. Several copies of genes of every step in
the fatty acid oxidation pathway, including acyl-CoA dehydroge-
nase, enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase,
and acetyl-CoA C-acetyltransferase, had ∼9-fold higher expres-
sion levels when prey, either bacteria or ciliates, were present
(Figures 3, 4A). Two genes encoding an electron transfer flavo-
protein gene and an electron transfer flavoprotein ubiquinone
oxidoreductase that also play a role in fatty acid oxidation (Wat-
mough and Frerman, 2010) had similar transcription patterns
(Table S2).
Fatty acid oxidation produces a large quantity of acetyl-
CoA, which can be used in two different pathways in P.
parvum mitochondria: complete oxidation through the TCA
cycle or transformation to succinate through the glyoxylate
cycle, with the former pathway being catabolic and the lat-
ter being anabolic. As expected, genes involved in both path-
ways had higher expression levels as a result of increased influx
of acetyl-CoA when prey were available (Figures 4B,C). Genes
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
FIGURE 3 | Relative expression levels of individual genes of P. parvum
within different metabolic pathways or functions. Each dot represents an
individual gene. X-axis values indicate expression levels in bacterized
treatment relative to those in axenic treatment (FPKMbacterized /FPKMaxenic ).
Y-axis values indicate expression levels in ciliate treatment relative to those in
axenic treatment (FPKMciliate /FPKMaxenic ). Sizes of the dots are proportional
to the number of reads assigned to the genes. Not all genes within certain
pathways or functions are shown, only those that have differential expression
between at least a pair of treatments. These functions and pathways with
significantly differential gene expression were plotted in Figures 3–5 to further
show consistent patterns within each function/pathway that are highlighted in
our discussion of core metabolism of P. parvum. For list of genes and their
read counts, FPKM values, and relative expression levels, refer to Table S2.
responsible for almost every step of the TCA cycle were reg-
ulated in strikingly similar patterns as those of fatty acid oxi-
dation (Figure 4B). Expression levels of two genes specifically
involved in the glyoxylate cycle, malate synthase and isocitrate
lyase (the rest of the pathway overlaps with the TCA cycle)
increased ∼6-fold in the bacterized treatment relative to the
axenic one, and increased ∼100-fold when ciliates were present
(Figure 4C).
Expression patterns of glycolysis/gluconeogenesis genes were
inconsistent among themselves. However, genes involved in the
connection between glycolysis/gluconeogenesis and other car-
bon pathways had interesting expression patterns. Genes encod-
ing pyruvate carboxylase and PEP carboxykinase had slightly
higher expression levels in the bacterized treatment, and even
higher levels in the ciliate treatment. These two proteins con-
vert pyruvate and TCA cycle metabolites through oxaloac-
etate to PEP, and are generally regarded as gluconeogenic
enzymes. Conversely, PEP carboxylase, which operates in the
opposite direction of PEP carboxykinase, decreased significantly
in its transcription level in the ciliate treatment (Figure 5,
Table S2).
Nitrogen Metabolism
Transcription data suggested dramatic changes in modes of
nitrogen acquisition in response to the addition of prey and par-
ticularly to the presence of the ciliates. First, expression levels
5 April 2015 | Volume 6 | Article 319
Liu et al.
FIGURE 4 | Average expression levels among different groups of
genes of P. parvum in bacterized and ciliate treatment relative to
those in axenic treatment. Geometric means ± 95% confidence intervals
of ammonium uptake genes drastically decreased in the ciliate
treatment, compared to the other two treatments (Figure 4D).
Those genes include several copies of an ammonium trans-
porter and two periplasmic L-amino acid oxidases. The lat-
ter two genes were similar to a homolog in the green algal
species, Chlamydomonas reinhardtii, that has been shown to
oxidize extracellular amino acids to release ammonium (Val-
lon et al., 1993). Similar although less dramatic transcriptional
changes were observed for nitrate reduction genes. Nitrate trans-
porter, nitrate reductase, ferredoxin-nitrite reductase and several
nitrite transporters had ∼6-fold lower expression levels in the
ciliate treatment relative to the axenic or bacterized treatment
(Figure 4E).
The balance and interconversion between α-ketoglutarate and
glutamate is central to nitrogen metabolism, especially amino
acid metabolism, as the pair provide the amino group to most
if not all amino acids during their biosynthesis, and accept
the amino group when they are metabolized. Expression lev-
els of glutamine synthetase (GS) and glutamine oxoglutarate
aminotransferase (GOGAT), which converts α-ketoglutarate to
glutamate, decreased significantly in the ciliate treatment. On
the other hand, expression levels of glutamate dehydrogenase
(GDH), which catalyzes the opposite reaction, increased sig-
nificantly (Figure 5, Table S2). Meanwhile, an aminotransferase
gene, which catalyzes the exchange of amino groups between
α-ketoglutarate/glutamate and other amino acids/α-ketoacids,
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
are shown for relative expression levels against the axenic treatment of genes
of each pathway or function. For list of genes and their read counts, FPKM
values, and relative expression levels, please refer to Table S2.
had significantly higher expression levels in the ciliate treatment
(Figure 5, Table S2).
Photosynthesis
A large number of genes involved in photosynthesis, such as most
photosynthetic reaction center genes and ribulose-bisphosphate
carboxylase genes are encoded by the chloroplast genome. Their
mRNAs were not retained in the polyA selection process in our
protocol, and their sequences were largely undetected in our
datasets. However, we were still able to obtain transcriptional pat-
terns of many chromosomal genes involved in photosynthesis or
CO2 fixation (Figure 4F). Four genes encoding photosystem I or
II proteins were detected in the transcriptome, and all of them
had higher expression levels when prey were present. The tran-
scriptome also included many copies of chlorophyll a/b binding
proteins, some of which were among the most actively tran-
scribed genes. Of these, 20 genes were differentially expressed,
and 17 of those had higher expression levels in the presence
of prey. Two genes encoding plant type ferredoxin, the major
electron acceptor from photosystem I, increased their expression
levels in the presence of prey. Three genes encoding carbonic
anhydrases, which supply CO2 to the Calvin cycle, had similar
expression patterns. Two other genes encoding phosphoglycerate
transporters that presumably transport the end products of the
Calvin cycle outside of chloroplast (Neuhaus and Wagner, 2000)
had similar expression patterns (Figure 5, Table S2).
6 April 2015 | Volume 6 | Article 319
Liu et al.
FIGURE 5 | Cellular overview of gene transcription of Prymnesium
parvum under three treatments and hypothetical model of metabolic
activities. Transcription levels of select genes, pathways or functions in
bacterized and ciliate treatments relative to axenic treatment are shown. Blue
shaded arrows and crosses indicate proposed model of metabolic flows and
Other Notable Pathways and Functions
Twenty-seven genes encoding ribosomal proteins exhibited sim-
ilar transcriptional patterns. On average, their expression lev-
els decreased ∼4-fold in the ciliate treatment compared to the
axenic treatment (Figure 4G). The same genes had slightly higher
expression levels (∼2 fold) in the bacterized treatment compared
to the axenic treatment.
Genes encoding three proteins involved in iron uptake were
regulated similarly. They include the low iron inducible periplas-
mic protein FEA1 and two high affinity iron permeases FTR1.
FEA1 has been shown to be one of the major proteins excreted
by iron deficient Chlamydomonas reinhardtii (Allen et al., 2007).
FEA1 may also facilitate iron uptake when expressed heterolo-
gously in yeast and plants (Narayanan et al., 2011). FTR1 is part
of a high affinity iron uptake system consisting of a permease and
an oxidase (Stearman et al., 1996). The transcriptional levels of
these three genes were ∼30-fold lower in the bacterized treat-
ment than in the axenic treatment. Those levels also decreased,
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
suppression of functions of P. parvum, respectively, in response to the
presence of bacteria in the bacterized treatment. Red shaded arrows and
crosses indicate the same flows and suppressions in response to the
presence of ciliates in the ciliate treatment. Cellular localizations of proteins
are based on knowledge of homologs in other protists or in silico predictions.
although to a lesser degree (∼8-fold) in the ciliate treatment
(Figure 4H).
Expression levels of 14 genes involving in purine biosynthe-
sis increased in both bacterized and ciliate treatments compared
to axenic treatment (Figure 4I). Most of the 15 genes encode
proteins involved in the early, generic part of the purine biosyn-
thesis pathway. Therefore, it was unclear whether there was a
targeted increase in adenine or guanine biosynthesis. In compar-
ison, pyrimidine biosynthesis genes did not change significantly
among the three treatments.
Genes responsible for the biosynthesis of prymnesins are
not yet fully characterized. However, since the two identified
prymnesins are polyketides (Igarashi et al., 1996), it has been
speculated that polyketide synthases play an important role in
the synthesis of prymnesins (Manning and La Claire, 2010).
Eleven different genes annotated as polyketide synthase were
found in the transcriptome. Among them, only one had differen-
tial expression (slightly higher in the ciliate treatment), the other
7 April 2015 | Volume 6 | Article 319
Liu et al.
10 were not differentially expressed in any of the treatments.
Homologs of the genes previously studied by qPCR (Freitag et al.,
2011) were among those not differentially expressed.
Three different phosphate transporters were differentially
expressed. However, the pattern of their expression among the
three treatments was not consistent. An ABC-type and a Pho4
family phosphate transporter both had higher expression levels
(∼4-fold) in the ciliate treatment. A sodium-dependent phos-
phate transporter had higher expression levels in the bacterized
treatment (∼4-fold).
Discussion
Comparison with Previous Gene Expression
Studies
The size of the P. parvum transcriptome is comparable
to other protistan transcriptomes in the Marine Microbial
Eukaryote Transcriptome Sequencing Project (https://www.
marinemicroeukaryotes.org), the transcriptomes of other prym-
nesiophytes (Koid et al., 2014), and protistan genomes in gen-
eral (Caron et al., 2009). The transcriptome was also compared
to EST data previously obtained for P. parvum. Out of 6286
EST sequences generated for P. parvum strain UTEX no. 2797
(La Claire, 2006), 5626 (89.5%) had a homolog in the transcrip-
tome of clone UOBS-LP0109 obtained in this study, with aver-
age nucleotide sequence identity of 98.9%. Out of 17,153 EST
sequences generated for P. parvum strain RL10 (Beszteri et al.,
2012), 14,271 (83.2%) had a homolog in this transcriptome, with
average nucleotide sequence identity of 95.8%. The numbers of
shared genes were probably slight underestimates because of the
limitations of the de novo short read assembly. P. parvum strain
UOBS-LP0109 shares a large numbers of genes with both other
strains, but was more closely related to strain UTEX no. 2797
than strain RL10 in terms of both percentage of shared genes and
sequence similarity. This finding is not surprising considering the
proximity of the sources of strain UOBS-LP0109 (Lake Texoma,
on the border of Oklahoma and Texas) and strain UTEX no. 2797
(Texas). Many of the most frequently-encountered ESTs in the
study of strain UTEX no. 2797 (La Claire, 2006), such as multiple
chlorophyll-binding proteins, ABC-type phosphate transporter,
and phosphoenolpyruvate carboxykinase, were also among the
genes with the most read pairs.
Nutrient Uptake from Prey
Numerous studies have documented the ability of P. parvum to
kill or prey on a variety organisms, ranging from bacteria to pho-
totrophic and heterotrophic protists, zooplankton and even fish
(Tillmann, 2003; Southard et al., 2010; Remmel and Hambright,
2012). It has been reported that heterotrophic nutrition by this
toxic, mixotrophic alga may provide a substantial percentage of
the macronutrients and carbon required for growth (Carvalho
and Granéli, 2010). However, the specific contribution to algal
nutrition provided by consumed prey remains unclear, and there
is virtually no information on how P. parvum might respond
at the cellular or molecular level to the availability of prey. We
evaluated changes in gene transcription of the alga in response
to two different prey assemblages: a mixed bacterial assemblage,
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
or one consisting of largely a ciliated protist with a background
assemblage of bacteria. We realize the lack of replication in
our transcriptome study and therefore have been purposefully
conservative in our interpretation of the data. In many cases,
different genes involved in the same pathway had similar
transcription patterns (Figure 3), thereby providing a level of
confidence to their collective transcription patterns. Inference
from the data was only drawn in those cases, and was never based
on the transcription pattern of a single gene. We also recognize
and acknowledge the complexity of the ciliate treatment in that
many factors could have affected gene transcription of P. parvum
leading to difficulties in attributing P. parvum response to a
specific factor. Those factors include ciliate cells, bacterial cells,
nutrients from the ciliate culture, and nutrients released as a
consequence of ciliates consuming bacteria. We assume that
P. parvum gene transcription changes in the ciliate treatment
were likely the result of its predation on ciliates because ciliate
biomass greatly outweighed bacterial biomass in that treatment,
but dissolved substances released by the ciliates may also have
elicited responses from the alga. Regardless, our results clearly
indicated that P. parvum killed prey and acquired nutrients in
their presence, and also provided insights into the response of
specific metabolic processes affected by the presence of other
microorganisms.
The transcriptomic data obtained in this study indicate some
shared changes in gene expression to the presence of bacte-
ria or ciliates (relative to the alga grown axenically), as well
as some changes that were specific to these two types of prey
assemblages. For carbon metabolism, expression levels of genes
involved in fatty acid metabolism and the TCA cycle were higher
in both treatments with prey (Figures 4A,B). This finding indi-
cates that P. parvum probably used fatty acids from both prey
types. Changes in expression levels of genes involved in the gly-
oxylate cycle, pyruvate carboxylase, PEP carboxykinase, and PEP
carboxylase were much more dramatic in the ciliate treatment
than in the bacterized treatment (Figure 5). This implies that
organic carbon from prey was used in different ways in those
two treatments. Specifically, it seemed that more fatty acids were
used for biomass production and less for energy production in
the ciliate treatment (Figure 5).
In nitrogen metabolism, significant changes in gene expres-
sion were only observed in the ciliate treatment. The presence
of ciliates led to dramatic decreases in expression levels of genes
involved in inorganic nitrogen uptake, especially ammonium
uptake (Figures 3, 4D). On the other hand, the transcription
patterns of GS, GOGAT, GDH and an aminotransferase gene
indicate increased catabolism of amino acids in the ciliate treat-
ment (Figure 5). Collectively, these data suggests that amino
acids, presumably obtained either directly from the ciliate or
mediated by the ciliate (i.e., perhaps as a consequence of cil-
iate predation on co-occurring bacteria), were metabolized in
P. parvum. Nitrogen needs of P. parvum were likely satisfied
by the influx of organic nitrogen, which led to the suppression
of inorganic nitrogen uptake. In all, these data imply that cil-
iates might be a very effective nitrogen source for P. parvum
in nature. Ciliates have been shown to be an important source
of nitrogen for other mixotrophic protists in nitrogen-limited
8 April 2015 | Volume 6 | Article 319
Liu et al.
environments (Bockstahler and Coats, 1993; Smalley and Coats,
2002). This hypothesis is also consistent with previous findings
that have reported greater toxicity of P. parvum under nitrogen
or phosphate limitation (Johansson and Granéli, 1999; Granéli
and Johansson, 2003; Hambright et al., 2014), and suppression
of toxicity when high doses of ammonium are provided (Grover
et al., 2007, 2013).
Expression levels of genes involved in iron uptake decreased
sharply in the presence of bacteria, and to a lesser degree in the
ciliate treatment (Figure 4H). Both prey assemblages were likely
sources of iron for P. parvum in our experiment, although it
would appear that bacteria alone were more effective in making
iron accessible than the ciliate. This finding is consistent with the
varied and complex relationships known to occur between phyto-
plankton and co-occurring bacteria (Amin et al., 2012). Maranger
et al. (1998) experimentally demonstrated that the ingestion of
bacteria by a different mixotrophic phytoflagellate, Ochromonas
sp., provided iron for cellular growth of the alga. We speculate
that the digestion of ingested prey by P. parvum may meet the
iron requirement for growth of the alga and thereby reduce the
need for iron uptake and transport. Alternatively, iron availabil-
ity may be increased in the presence of other microbes, obviating
a requirement for phagotrophy by P. parvum. For example, pre-
dation on bacteria by heterotrophic protists has been shown to
relieve iron stress of co-occurring algae (Barbeau et al., 1996).
Most previous studies of P. parvum have focused on the impor-
tance of nitrogen and phosphate contained in prey biomass as
a factor in the toxicity of the alga, since N:P ratio has a direct
impact on the toxicity of P. parvum (Granéli et al., 2012). Our
results broaden these findings to include the possible importance
of iron to the metabolism of P. parvum. The bacterial assem-
blage in our culture was part of the native bacterial assemblage
occurring during a P. parvum bloom. It would be interesting to
determine whether P. parvum has a preference among different
bacteria as a means of acquiring iron.
Contribution of Predation to the Growth of
P. parvum
The ciliate employed in the study, Uronema marina, was readily
killed when introduced into cultures of P. parvum at high abun-
dance. The process involved immobilization and attachment of
several algal cells to each ciliate, as has been described previ-
ously (Tillmann, 2003). Bacterivory has also been reported for
P. parvum (Nygaard and Tobiesen, 1993), although ingestion of
bacteria was not confirmed in our study. It has been speculated
that prey capture and digestion by P. parvum provide substantial
sources of major nutrients (N, P) or organic carbon for energy
and growth (Carvalho and Granéli, 2010; Granéli et al., 2012).
Our transcriptional data supported this hypothesis, for nitrogen
and carbon.
Many mixotrophic protists grow faster when prey are avail-
able compared to axenic growth (Sanders, 2011). Surprisingly,
our transcriptional data and our direct measurements of growth
of the alga indicated that P. parvum did not grow faster despite its
demonstrable predatory behavior. Protein synthesis at ribosomes
is a huge component of cell growth, therefore ribosome synthesis
could at least in part reflect cell growth. It has been proposed that
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
ribosome synthesis levels could be a good measure for growth
(Rudra and Warner, 2004). The expression levels of genes encod-
ing P. parvum ribosomal proteins increased only slightly in the
bacterized treatment compared to levels observed for the axenic
treatment (Figure 4G). However, P. parvum did not grow signif-
icantly faster in the presence of bacteria (Figure 1). On the other
hand, the expression levels of these genes decreased in the cili-
ate treatment (Figure 4G), implying P. parvum synthesized fewer
ribosomes. Although the addition of ciliates in the late exponen-
tial phase maintained growth of P. parvum, that growth rate was
slower than that of an axenic culture in the early-mid exponen-
tial phase (p < 0.001, ANOVA, Figure 1). The treatments with
prey did support slightly higher yields of the alga relative to the
axenic P. parvum culture. This outcome is presumably a result
of the additional nutrients available from the prey or the culture
medium of the prey. Overall, our findings are consistent with a
previous study in which P. parvum fed Rhodomonas salina did
not grow faster than a monoculture of P. parvum, even under
nutrient deficient conditions (Carvalho and Granéli, 2010).
One possible explanation for the lack of increased growth rate
in the presence of prey is the cost of making heterotrophic cellu-
lar machinery when prey become available. However, we did not
observe an initial decrease followed by later recovery in growth
rate that would be expected if such hypothesis were true. Addi-
tionally, killing of the ciliate cells started almost instantaneously
when they were added to the P. parvum culture, suggesting P.
parvum was already toxic and predatory when growing axeni-
cally. Therefore, a more likely explanation is that P. parvum is
critically dependent on some aspect of phototrophy, and that
phototrophy presents a bottleneck controlling growth rate, at
least under the specific growth conditions employed in this study.
A photosynthetic bottleneck is implied by the observation that
P. parvum cannot grow or survive in the dark even when prey
are available at high abundances (Carvalho and Granéli, 2010;
Brutemark and Granéli, 2011). The only known exception to this
rule is that P. parvum can grow very slowly in the dark when
supplied with high concentrations of certain organic compounds
(Rahat and Jahn, 1965). However, the concentrations of organic
compounds required for its growth in the dark are much higher
than concentrations that would be experienced in nature. The
mechanism and metabolic details of such growth are unknown.
Carvalho and Granéli (2010) suggested that photosynthesis
and phagotrophy in P. parvum might supplement each other,
employing “co-metabolism” for growth. Our results do not
directly support this hypothesis, since the presence of prey in our
study did not enhance growth rate per se beyond that enabled
by phototrophic growth. However, one observation in our study
was at least partially consistent with their hypothesis. A num-
ber of genes involved in photosynthesis, although far from com-
plete and conclusive, increased their expression levels in response
to prey despite the fact that there was no difference in light
conditions during the experiment. The presence of prey may
have relieved some of the cellular processes required to obtain
major nutrient elements (e.g., N), as evidenced by changes in
gene expression among our experimental treatments. As a con-
sequence, more resources might have been made available for
making the photosynthetic apparatus. Alternatively, prey may
9 April 2015 | Volume 6 | Article 319
Liu et al.
have provided some micronutrient(s) that support photosynthe-
sis. For example, increased availability of iron may have had a
positive effect on photosynthesis since the photosynthetic appa-
ratus includes many proteins and cofactors that contain Fe-S
clusters.
It was unclear from our transcriptomic data where or how
nutrients acquired by P. parvum from its heterotrophic activi-
ties were used, since growth rate was not noticeably enhanced.
One possibility is storage, possibly in carbohydrates. Some of our
data supported this hypothesis as expression levels of some glu-
coneogenic genes increased in treatments with prey, especially
in the ciliate treatment (Figure 5). However, genes involved in
starch biosynthesis were not significantly regulated. Other forms
of storage are also possible, although there were no transcrip-
tional data indicating which form. If this hypothesis is true, it
means that P. parvum may be acquiring nutrients from prey
and storing them for growth when conditions are amenable for
photosynthesis. This is consistent with the observation that P.
parvum took up nitrogen and phosphorus from Myrionecta rubra
but did not grow in the dark (Brutemark and Granéli, 2011).
Our study investigated expression of a wide range of genes in
the mixotrophic protist P. parvum in the presence or absence of
two different prey assemblages. Our gene expression data clearly
indicated that P. parvum obtained a variety of nutrients from its
prey including organic carbon in the form of fatty acids from
References
Allen, M. D., del Campo, J. A., Kropat, J., and Merchant, S. S. (2007). FEA1,
FEA2, and FRE1, encoding two homologous secreted proteins and a can-
didate ferrireductase, are expressed coordinately with FOX1 and FTR1 in
iron-deficient Chlamydomonas reinhardtii. Eukaryot. Cell 6, 1841–1852. doi:
10.1128/EC.00205-07
Amin, S. A., Parker, M. S., and Armbrust, E. V. (2012). Interactions
between diatoms and bacteria. Microbiol. Mol. Biol. Rev. 76, 667–684. doi:
10.1128/MMBR.00007-12
Barbeau, K., Moffett, J. W., Caron, D. A., Croot, P. L., and Erdner, D. L. (1996).
Role of protozoan grazing in relieving iron limitation of phytoplankton. Nature
380, 61–64. doi: 10.1038/380061a0
Barkoh, A., and Fries, L. T. (2010). Aspects of the origins, ecology, and control
of Golden Alga Prymnesium parvum: introduction to the featured collection.
J. Am. Water Resour. Assoc. 46, 1–5. doi: 10.1111/j.1752-1688.2009.00394.x
Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate:
a practical and powerful approach to multiple testing. J. R. Stat. Soc. B 57,
289–300.
Bertin, M. J., Zimba, P. V., Beauchesne, K. R., Huncik, K. M., and Moeller, P. D.
R. (2012). The contribution of fatty acid amides to Prymnesium parvum carter
toxicity. Harmful Algae 20, 117–125. doi: 10.1016/j.hal.2012.08.004
Beszteri, S., Yang, I., Jaeckisch, N., Tillman, U., Frickenhaus, S., Glöckner, G.,
et al. (2012). Transcriptomic response of the toxic prymnesiophyte Prymnesium
parvum (N. Carter) to phosphorus and nitrogen starvation. Harmful Algae 18,
1–15. doi: 10.1016/j.hal.2012.03.003
Bik, H. M., and Thomas, W. K. (2012). Metagenomics will highlight and drive
links to taxonomic data: reply to Murray. Trends Ecol. Evol. 27, 652–653. doi:
10.1016/j.tree.2012.09.005
Bockstahler, K. R., and Coats, D. W. (1993). Grazing of the mixotrophic dinoflagel-
late Gymnodinium sanguineum on ciliate populations of Chesapeake Bay. Mar.
Biol. 116, 477–487. doi: 10.1007/BF00350065
Brooks, B. W., James, S. V., Valenti, J. T. W., Urena-Boeck, F., Serrano, C.,
Berninger, J. P., et al. (2010). Comparative toxicity of Prymnesium parvum
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
both the bacteria and the ciliate, organic nitrogen in the form of
amino acids from the ciliate, and iron from the bacteria. Despite
the nutrient gains, P. parvum did not grow faster, implying that
its growth is ultimately controlled by its photosynthetic activity,
at least in the cases of tested prey types. Our study revealed tran-
scriptional behaviors of P. parvum in response to prey availability,
and has led to several testable hypotheses relating to interactions
between P. parvum and its prey. These results provide the neces-
sary foundation for future molecular studies regarding the con-
tribution of prey for growth and maintenance of this ecologically
and economically important organism.
Acknowledgments
This research is funded by the Gordon and Betty Moore Founda-
tion through Grant GBMF3299 to DAC and KBH. This sequenc-
ing effort was funded in part by the Gordon and Betty Moore
Foundation through Grant GBMF2637 to the National Center for
Genome Resources.
Supplementary Material
The Supplementary Material for this article can be found
online at: http://www.frontiersin.org/journal/10.3389/fmicb.
2015.00319/abstract
in Inland waters. J. Am. Water Resour. Assoc. 46, 45–62. doi: 10.1111/j.1752-
1688.2009.00390.x
Brutemark, A., and Granéli, E. (2011). Role of mixotrophy and light for growth
and survival of the toxic haptophyte Prymnesium parvum. Harmful Algae 10,
388–394. doi: 10.1016/j.hal.2011.01.005
Caron, D. A., Worden, A. Z., Countway, P. D., Demir, E., and Heidelberg, K.
B. (2009). Protists are microbes too: a perspective. ISME J. 3, 4–12. doi:
10.1038/ismej.2008.101
Carvalho, W. F., and Granéli, E. (2010). Contribution of phagotrophy
versus autotrophy to Prymnesium parvum growth under nitrogen and
phosphorus sufficiency and deficiency. Harmful Algae 9, 105–115. doi:
10.1016/j.hal.2009.08.007
Conesa, A., Götz, S., Garcia-Gomez, J. M., Terol, J., Talon, M., and Rob-
les, M. (2005). Blast2GO: a universal tool for annotation, visualization and
analysis in functional genomics research. Bioinformatics 21, 3674–3676. doi:
10.1093/bioinformatics/bti610
Edvardsen, B., and Paasche, E. (1998). “Bloom dynamics and physiology of
Prymnesium and Chrysochromulina,” in Physiological Ecology of Harmful Algal
Blooms, Vol. G 41, eds D. M. Anderson, A. D. Cembella, and G. M. Hallegraef
(Berlin: Springer-Verlag), 193–208.
Fistarol, G. O., Legrand, C., and Granéli, E. (2003). Allelopathic effect of Prym-
nesium parvum on a natural plankton community. Mar. Ecol. Prog. Ser. 255,
115–125. doi: 10.3354/meps255115
Freitag, M., Beszteri, B., Vogel, H., and John, U. (2011). Effects of physiologi-
cal shock treatments on toxicity and polyketide synthase gene expression in
Prymnesium parvum (Prymnesiophyceae). Eur. J. Phycol. 46, 193–201. doi:
10.1080/09670262.2011.591438
Granéli, E., Edvardsen, B., Roelke, D. L., and Hagström, J. A. (2012). The ecophys-
iology and bloom dynamics of Prymnesium spp. Harmful Algae 14, 260–270.
doi: 10.1016/j.hal.2011.10.024
Granéli, E., and Johansson, N. (2003). Increase in the production of allelo-
pathic substances by Prymnesium parvum cells grown under N- or P-
deficient conditions. Harmful Algae 2, 135–145. doi: 10.1016/S1568-9883(03)
00006-4
10 April 2015 | Volume 6 | Article 319
Liu et al.
Granéli, E., and Salomon, P. S. (2010). Factors influencing allelopathy and tox-
icity in Prymnesium parvum. J. Am. Water Resour. Assoc. 46, 108–120. doi:
10.1111/j.1752-1688.2009.00395.x
Grover, J. P., Baker, J. W., Ureña-Boeck, F., Brooks, B. W., Errera, R. M., Roelke,
D. L., et al. (2007). Laboratory tests of ammonium and barley straw extract as
agents to suppress abundance of the harmful alga Prymnesium parvum and its
toxicity to fish. Water Res. 41, 2503–2512. doi: 10.1016/j.watres.2007.03.025
Grover, J. P., Roelke, D. L., Brooks, B. W., Gable, G. M., Neisch, M. T.,
Hayden, N. J., et al. (2013). Ammonium treatments to suppress toxic
blooms of Prymnesium parvum in a subtropical lake of semi-arid climate:
results from in situ mesocosm experiments. Water Res. 47, 4274–4285. doi:
10.1016/j.watres.2013.05.001
Guo, M., Harrison, P. J., and Taylor, F. J. R. (1996). Fish kills related to Prymne-
sium parvum N. Carter (Haptophyta) in the People’s Republic of China. J. Appl.
Phycol. 8, 111–117. doi: 10.1007/BF02186313
Hambright, K. D., Beyer, J. E., Easton, J. D., Zamor, R. M., Easton, A. C., and Hall-
idayschult, T. C. (2015). The niche of an invasive marine microbe in a subtrop-
ical freshwater impoundment. ISME J. 9, 256–264. doi: 10.1038/ismej.2014.103
Hambright, K. D., Easton, J. D., Zamor, R. M., Beyer, J., Easton, A. C., and Allison,
B. (2014). Regulation of growth and toxicity of a mixotrophic microbe: impli-
cations for understanding range expansion in Prymnesium parvum. Freshwater
Sci. 33, 745–754. doi: 10.1086/677198
Henrikson, J. C., Gharfeh, M. S., Easton, A. C., Easton, J. D., Glenn, K. L., Shadfan,
M, et al. (2010). Reassessing the ichthyotoxin profile of cultured Prymnesium
parvum (golden algae) and comparing it to samples collected from recent fresh-
water bloom and fish kill events in North America. Toxicon 55, 1396–1404. doi:
10.1016/j.toxicon.2010.02.017
Huang, X., and Madan, A. (1999). CAP3: A DNA sequence assembly program.
Genome Res. 9, 868–877. doi: 10.1101/gr.9.9.868
Igarashi, T., Satake, M., and Yasumoto, T. (1996). Prymnesin-2: a potent ichthy-
otoxic and hemolytic glycoside isolated from the red tide alga Prymnesium
parvum. J. Am. Chem. Soc. 118, 479–480. doi: 10.1021/ja9534112
Iseli, C., Jongeneel, C. V., and Bucher, P. (1999). ESTScan: a program for detecting,
evaluating, and reconstructing potential coding regions in EST sequences. Proc.
Int. Conf. Intell. Syst. Mol. Biol. 138–148.
Johansson, N., and Granéli, E. (1999). Influence of different nutrient conditions on
cell density, chemical composition and toxicity of Prymnesium parvum (Hap-
tophyta) in semi-continuous cultures. J. Exp. Mar. Biol. Ecol. 239, 243–258. doi:
10.1016/S0022-0981(99)00048-9
Johnsen, T. M., Eikrem, W., Olseng, C. D., Tollefsen, K. E., and Bjerknes, V. (2010).
Prymnesium parvum: the Norwegian experience. J. Am. Water Resour. Assoc.
46, 6–13. doi: 10.1111/j.1752-1688.2009.00386.x
Jones, A. C., Liao, T. S. V., Najar, F. Z., Roe, B. A., Hambright, K. D.,
and Caron, D. A. (2013). Seasonality and disturbance: annual pattern and
response of the bacterial and microbial eukaryotic assemblages in a fresh-
water ecosystem. Environ. Microbiol. 15, 2557–2572. doi: 10.1111/1462-2920.
12151
Jones, H. (1997). A classification of mixotrophic protists based on their behaviour.
Freshwater Biol. 37, 35–43. doi: 10.1046/j.1365-2427.1997.00138.x
Keeling, P. J., Burki, F., Wilcox, H. M., Allam, B., Allen, E. E., Amaral-Zettler,
L. A., et al. (2014). The Marine Microbial Eukaryote Transcriptome Sequenc-
ing Project (MMETSP): Illuminating the functional diversity of eukaryotic life
in the oceans through transcriptome sequencing. PLoS Biol. 12:e1001889. doi:
10.1371/journal.pbio.1001889
Koid, A. E., Liu, Z., Terrado, R., Jones, A. C., Caron, D. A., and Heidelberg, K.
B. (2014). Comparative transcriptome analyses of four Prymnesiophyte algae.
PLoS ONE 9:e97801. doi: 10.1371/journal.pone.0097801
La Claire, J. W. II. (2006). Analysis of expressed sequence tags from the harmful
alga, Prymnesium parvum (Prymnesiophyceae, Haptophyceae). Mar. Biotech-
nol. 8, 534–546. doi: 10.1007/s10126-005-5182-2
Larsen, A. (1999). Prymnesium parvum and P. patelliferum (Haptophyta) – one
species. Phycologia 38, 541–543. doi: 10.2216/i0031-8884-38-6-541.1
Li, H., and Durbin, R. (2009). Fast and accurate short read alignment
with Burrows-Wheeler Transform. Bioinformatics 25, 1754–1760. doi:
10.1093/bioinformatics/btp324
Li, W., and Godzik, A. (2006). Cd-hit: a fast program for clustering and comparing
large sets of protein or nucleotide sequences. Bioinformatics 22, 1658–1659. doi:
10.1093/bioinformatics/btl158
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
Luo, R., Liu, B., Xie, Y., Li, Z., Huang, W., Yuan, J., et al. (2012). SOAPden-
ovo2: an empirically improved memory-efficient short-read de novo assembler.
Gigascience 1:18. doi: 10.1186/2047-217X-1-18
Lutz-Carrillo, D. J., Southard, G. M., and Fries, L. T. (2010). Global genetic rela-
tionships among isolates of Golden Alga (Prymnesium parvum). J. Am. Water
Resour. Assoc. 46, 24–32. doi: 10.1111/j.1752-1688.2009.00388.x
Manning, S. R., and La Claire, J. W. II. (2010). Prymnesins: toxic metabolites
of the Golden Alga, Prymnesium parvum Carter (Haptophyta). Mar. Drugs 8,
678–704. doi: 10.3390/md8030678
Manning, S. R., and La Claire, J. W. II. (2013). Isolation of polyketides from
Prymnesium parvum (Haptophyta) and their detection by liquid chromatog-
raphy/mass spectrometry metabolic fingerprint analysis. Anal. Biochem. 442,
189–195. doi: 10.1016/j.ab.2013.07.034
Maranger, R., Bird, D. F., and Price, N. M. (1998). Iron acquisition by photosyn-
thetic marine phytoplankton from ingested bacteria. Nature 396, 248–251. doi:
10.1038/24352
McLaughlin, J. J. A. (1958). Euryhaline chrysomonads: nutrition and toxigenesis in
Prymnesium parvum, with notes on Isochrysis galbana and Monochrysis lutheri.
J. Protozool. 5, 75–81. doi: 10.1111/j.1550-7408.1958.tb02529.x
Murray, S. A., Patterson, D. J., and Thessen, A. E. (2012). Transcriptomics and
microbial eukaryote diversity: a way forward. Trends. Ecol. Evol. 27, 651–652.
doi: 10.1016/j.tree.2012.08.011
Narayanan, N. N., Ihemere, U., Chiu, W. T., Siritunga, D., Rajamani, S., Singh, S.,
et al. (2011). The iron assimilatory protein, FEA1, from Chlamydomonas rein-
hardtii facilitates iron-specific metal uptake in yeast and plants. Front. Plant,
Sci. 2:67. doi: 10.3389/fpls.2011.00067
Neuhaus, H. E., and Wagner, R. (2000). Solute pores, ion channels, and metabo-
lite transporters in the outer and inner envelope membranes of higher
plant plastids. Biochim. Biophys. Acta. 1465, 307–323. doi: 10.1016/S0005-
2736(00)00146-2
Nygaard, K., and Tobiesen, A. (1993). Bacterivory in algae: a survival
strategy during nutrient limitation. Limnol. Oceanogr. 38, 273–279. doi:
10.4319/lo.1993.38.2.0273
Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., et al. (2013). The
SILVA ribosomal RNA gene database project: improved data processing and
web-based tools. Nucl. Acids Res. 41, D590–D596. doi: 10.1093/nar/gks1219
Rahat, M., and Jahn, T. L. (1965). Growth of Prymnesium parvum in the dark:
note on ichthyotoxin formation. J. Protozool. 12, 246–250. doi: 10.1111/j.1550-
7408.1965.tb01845.x
Rahat, M., and Reich, K. (1963). The B12 vitamins and growth of the flagellate
Prymnesium parvum. J. Gen. Microbiol. 31, 195–202. doi: 10.1099/00221287-
31-2-195
Remmel, E. J., and Hambright, K. D. (2012). Toxin-assisted micropredation: exper-
imental evidence shows that contact micropredation rather than exotoxicity
is the role of Prymnesium toxins. Ecol. Lett. 15, 126–132. doi: 10.1111/j.1461-
0248.2011.01718.x
Remmel, E. J., Kohmescher, N., Larson, J. H., and Hambright, K. D. (2011). An
experimental analysis of harmful algae-zooplankton interactions and the ulti-
mate defense. Limnol. Oceanogr. 56, 461–470. doi: 10.4319/lo.2011.56.2.0461
Robertson, G., Schein, J., Chiu, R., Corbett, R., Field, M., Jackman, S. D., et al.
(2010). De novo assembly and analysis of RNA-seq data. Nat. Methods 7,
909–912. doi: 10.1038/nmeth.1517
Robinson, M. D., McCarthy, D. J., and Smyth, G. K. (2010). edgeR: a Bioconduc-
tor package for differential expression analysis of digital gene expression data.
Bioinformatics 26, 139–140. doi: 10.1093/bioinformatics/btp616
Roelke, D. L., Grover, J. P., Brooks, B. W., Glass, J., Buzan, D., Southard, G. M.,
et al. (2011). A decade of fish-killing Prymnesium parvum blooms in Texas:
roles of inflow and salinity. J. Plankton Res. 33, 243–253. doi: 10.1093/plankt/
fbq079
Rudra, D., and Warner, J. R. (2004). What better measure than ribosome synthesis?
Genes Dev. 18, 2431–2436. doi: 10.1101/gad.1256704
Sanders, R. W. (2011). Alternative nutritional strategies in protists: symposium
introduction and a review of freshwater protists that combine photosynthe-
sis and heterotrophy. J. Eukaryot. Microbiol. 58, 181–184. doi: 10.1111/j.1550-
7408.2011.00543.x
Schmidt, L. E, and Hansen, P. J. (2001). Allelopathy in the prymnesiophyte
Chrysochromulina polylepis: effect of cell concentration, growth phase and pH.
Mar. Ecol. Prog. Ser. 216, 67–81. doi: 10.3354/meps216067
11 April 2015 | Volume 6 | Article 319
Liu et al.
Schug, K. A., Skingel, T. R., Spencer, S. E., Serrano, C. A., Le, C. Q., Schug, C. A.,
et al. (2010). Hemolysis, fish mortality, and LC-ESI-MS of cultured crude and
fractionated Golden Alga (Prymnesium parvum). J. Am. Water Resour. Assoc.
46, 33–44. doi: 10.1111/j.1752-1688.2009.00389.x
Simpson, J. T., Wong, K., Jackman, S. D., Schein, J. E., Jones, S. J., and Birol, I.
(2009). ABySS: a parallel assembler for short read sequence data. Genome Res.
19, 1117–1123. doi: 10.1101/gr.089532.108
Smalley, G. W., and Coats, D. W. (2002). Ecology of the red-tide dinoflagel-
late Ceratium furca: distribution, mixotrophy, and grazing impact on cili-
ate populations of Chesapeake Bay. J. Eukaryot. Microbiol. 49, 63–73. doi:
10.1111/j.1550-7408.2002.tb00343.x
Sopanen, S., Koski, M., Kuuppo, P., Uronen, P., Legrand, C., and Tamminen,
T. (2006). Toxic haptophyte Prymnesium parvum affects grazing, survival,
egestion and egg production of the calanoid copepods Eurytemora affinis
and Acartia bifilosa. Mar. Ecol. Progr. Ser. 327, 223–232. doi: 10.3354/meps
327223
Southard, G. M., Fries, L. T., and Barkoh, A. (2010). Prymnesium parvum: the
Texas experience. J. Am. Water Resour. Assoc. 46, 14–23. doi: 10.1111/j.1752-
1688.2009.00387.x
Stearman, R., Yuan, D. S., Yamaguchi-Iwai, Y., Klausner, R. D., and Dancis, A.
(1996). A permease-oxidase complex involved in high-affinity iron uptake in
yeast. Science 271, 1552–1557. doi: 10.1126/science.271.5255.1552
Sunda, W. G., Granéli, E., and Gobler, C. J. (2006). Positive feedback and the devel-
opment and persistence of ecosystem disruptive algal blooms. J. Phycol. 42,
963–974. doi: 10.1111/j.1529-8817.2006.00261.x
Frontiers in Microbiology | www.frontiersin.org
Prymnesium parvum transcriptome and prey
Tillmann, U. (2003). Kill and eat your predator: a winning strategy of the
planktonic flagellate Prymnesium parvum. Aq. Microb. Ecol. 32, 73–84. doi:
10.3354/ame032073
Vallon, O., Bulté, L., Kuras, R., Olive, J., and Wollman, F. A. (1993). Exten-
sive accumulation of an extracellular L-amino-acid oxidase during gameto-
genesis of Chlamydomonas reinhardtii. Eur. J. Biochem. 215, 351–360. doi:
10.1111/j.1432-1033.1993.tb18041.x
Watmough, N. J., and Frerman, F. E. (2010). The electron transfer flavoprotein:
ubiquinone oxidoreductaases. Biochim. Biophys. Acta 1797, 1910–1916. doi:
10.1016/j.bbabio.2010.10.007
Weissbach, A., and Legrand, C. (2012). Effect of different salinities on growth and
intra- and extracellular toxicity of four strains of the haptophyte Prymnesium
parvum. Aq. Microb. Ecol. 67, 139–149. doi: 10.3354/ame01589
Conflict of Interest Statement: The authors declare that the research was con-
ducted in the absence of any commercial or financial relationships that could be
construed as a potential conflict of interest.
Copyright © 2015 Liu, Jones, Campbell, Hambright, Heidelberg and Caron. This is
an open-access article distributed under the terms of the Creative Commons Attribu-
tion License (CC BY). The use, distribution or reproduction in other forums is per-
mitted, provided the original author(s) or licensor are credited and that the original
publication in this journal is cited, in accordance with accepted academic practice.
No use, distribution or reproduction is permitted which does not comply with these
terms.
12 April 2015 | Volume 6 | Article 319